*Do-file for the research article:
*Interpreting expectations: Normative and predictive expectations as seen by citizens
*Morten Hjortskov (2020)
*Journal of Behavioral Public Administration

*Data: interpreting_expectations.dta

***************************************************************************************************
*Study 1: Interpretations of the word expectations and predictive and the normative question (Experiment 4)
***************************************************************************************************

*******************
*Figure 1 - Interpretation of the Word “Expectations”
*******************

tab intexp


*******************
*Figure 2 - Citizens’ Interpretation of Two Common Questions about Expectations
*******************

tab intpos
tab intnorm



***************************************************************************************************
*Study 2: Presenting both the predictive and the normative question 
***************************************************************************************************

*******************
*Table 3 - Average Expectations across Experimental Groups
*******************

table exp3, c(mean pos mean norm)


*******************
*Table 4 - Test of Differences between Experimental Groups
*******************

*Generating combined expectations variable
gen posnorm = pos
replace posnorm = norm if exp3==2 | exp3==5


*Effect of Hometown treatment on positive and normative expectations (Single questions)
eststo clear
regress pos highexp if poscond==1, vce(robust) //model 1
eststo

regress norm highexp if normcond==1, vce(robust) //model 2
eststo


*Effect of Hometown treatment on positive and normative expectations (Both questions)
regress pos highexp if bothcond==1, vce(robust) //model 3
eststo

regress norm highexp if bothcond==1, vce(robust) //model 4
eststo

*Difference in positive expectations between positive and both conditions
regress pos bothcond if lowexp==1, vce(robust) //model 5
eststo
esize twosample pos if lowexp==1, by(bothcond) //Cohen's d

regress pos bothcond if highexp==1, vce(robust) //model 6
eststo

esttab using study2.rtf, se replace r2 compress nogaps label star(+ .1 * .05 ** .01)
